Exploring Neuro-AI: Merging Neuroscience w/ Generative AI for Innovation & Patents w/ A. K. Pradeep

[00:00:02] So we're talking about AI and generative AI. But what if we could layer neuroscience into it? What is neuro-AI, and how might it affect the outcomes? I get to talk to Dr. A. K. Pradeep, who's a pioneer in this space, and he walks through how the technology works, how you can use it for innovation and new ideas, and how it can even help you with the patents.

[00:00:25] It's fantastic for small businesses, and we learn more on this bonus episode, The Business of Tech.

[00:00:33] Thread has declared death to the ticket. Did you know tickets date back to the 1800s? And customers hate tickets. Thread uses an approach for connecting, communicating, and ultimately, collaborating.

[00:00:47] Thread allows people to come together around a topic. They can discuss and make decisions. They can share and invite the right people.

[00:00:55] The future of service is collaborative. Supercharge your service experience by seamlessly integrating AI and automation to meet your customers where they are, in Teams, Slack, and desktops, not in tickets.

[00:01:09] With instant updates to and from ConnectWise, Autotask, and Halo PSA, keep all communication in one place, where it should be, with people.

[00:01:18] Decrease time to resolution 30% with chat-based support by visiting getthread.com slash MSPRadio to declare death to the ticket.

[00:01:31] Well, Dr. Pradeep, welcome to the show.

[00:01:34] Dave, thank you very much for having me here. It's a pleasure.

[00:01:39] Dave, thank you very much for having me here.

[00:02:07] Thank you very much.

[00:02:24] work done by neuroscientists, people in market research, everything. Now, it is a pity and a

[00:02:32] crying shame if we took all those learnings, bundled them, and threw them away and said,

[00:02:39] Chad GPT, tell me what I should do. It seems to be ridiculous. So what is wonderful is we could

[00:02:46] have building blocks. Think of Lego blocks. Gen AI is one Lego block. We honor the other Lego blocks

[00:02:54] and connect them together to build something powerful. So now, the same GPT algorithms,

[00:03:02] transformers, and Chad GPTs, when powered through the learnings of what motivates consumers,

[00:03:09] what persuades them and seduces them into buying and doing things, if you could blend it

[00:03:15] with generative AI, now it becomes useful for business. Otherwise, it's cute pros, but not

[00:03:23] particularly useful for business. So it gives consumers and marketeers confidence that whatever

[00:03:31] their algorithms come up with, they're not just informed by creativity, but they're informed through

[00:03:40] an understanding of human behavior as manifested through the progress in neuroscience. So bringing

[00:03:47] the lifeblood of commerce back into the algorithms of creativity is what we call neural AI. It is

[00:03:57] neuroscience offering generative AI.

[00:04:02] Okay, that sounds really clever. And I want to make sure that I understand how it works. Because,

[00:04:06] you know, as one who's been playing with the models, we have the base training models, and then we can

[00:04:09] teach it more. We can make it an expert in our own business, our own customer base by giving it

[00:04:16] additional training data as it works. Is this as simple as adding neuroscience training data to it? Or is it

[00:04:24] introducing another kind of learning to the back end? Help me walk through how this actually works in

[00:04:29] practice. So when you think about it, right, when you think how transformers that are trained work

[00:04:34] really well, the transformer produces an output. Then typically, a human being looks at the output and

[00:04:41] says, oh, that's kind of cool. And that's uncool. So the transformer says, oh, I will learn from what

[00:04:46] you found to be cool and uncool and trains itself. That's how the algorithms of today work. Now, when you

[00:04:53] think about it, you say, my God, I need to persuade a teenager. And the algorithm comes up with some pros,

[00:05:01] and then says, this is going to persuade a teenager. However, I don't know if you have teens, teen children,

[00:05:07] Dave, but you will quickly realize, never reason with a teenager. There's a learning for the teen brain.

[00:05:13] The teen brain processes everything not through the lens of reason, but through the lens of emotion.

[00:05:18] Right? It's a neuroscience learning. Do we have to wait for 10,000 paragraphs to be generated by the

[00:05:25] algorithm, and people say yay or nay for the algorithm to learn? Or can we say, excuse me,

[00:05:32] thousands of experiments have been run, and there are learnings from what it is that persuades the teen

[00:05:38] brain? As an example, use the language of emotion, not of reason. Can the algorithm be rethought

[00:05:45] with that learning? Can these rules of neuroscience be embedded smartly in an algorithmic framework so

[00:05:54] that we don't have to go through the endless, tedious process of output correction learning? So the idea is

[00:06:02] that sometimes output correction and learning are good. Sometimes I don't need to know to touch a hot

[00:06:10] rod is going to be injurious for my skin. Excuse me, I could tell the algorithm that. So the idea of taking

[00:06:18] neuroscience insights and learnings and embedding them as part of an algorithmic framework is really

[00:06:27] how it is done. That's very clever. And I like the combination of that. Now, we both know that most

[00:06:35] organizations are leveraging AI-based technology, be it generative AI, or if now we're trying to talk

[00:06:42] about neuro AI here. They're generally leveraging that by off-the-shelf products. I don't have an

[00:06:48] expectation that the majority of businesses around the world are building and training their own model.

[00:06:54] At best, they're taking a base model of some kind and then giving it access to the data. Now,

[00:07:01] I'm going to put aside for a moment the governance and data prep stuff that has to happen. Let's assume

[00:07:07] that's already happened. And for listeners, you've had plenty of interviews talking about how that's

[00:07:11] done. So let's assume that's done reasonably well. At this point, how do I leverage a neuro AI in this

[00:07:18] fund? Is it a matter of building the right new training data for my targets or is it combining it?

[00:07:24] Help me understand how I make this practical.

[00:07:27] That's a great question. And what we have worked hard is that nobody has to do this.

[00:07:32] Chat GPT or OpenAI doesn't come to you and say, you know, Dave, you should train your own

[00:07:39] podcasting transformer so you don't have to show up anymore, right? It doesn't do that,

[00:07:44] right? And thank God, I hope it never will, right? But the idea is our platforms that we create in my

[00:07:51] company, sensory.ai, are fully trained already with the neuroscience embedded in it. Because

[00:07:58] our target, as I mentioned to you, Dave, in our pre-meeting, our focus is small and medium-sized

[00:08:04] businesses, right? Even though the large companies use my platform, nobody has time to come and do

[00:08:11] any training. You may be selling soup or you may be creating power bars or you may be creating

[00:08:17] cookies. You are just have enough time to manage your business. You do not have the time to go train

[00:08:24] a model into creating messaging or packaging or anything for you. So what we do is the neuroscience

[00:08:33] learnings, and again, they go by, you know, cisgendered male brain, cisgendered female brains,

[00:08:41] brains between 40 and 50, teen brains, brains over 50 to 60, above 60, all of them, the neural rules are

[00:08:51] already embedded. You don't have to go reinvent a wheel, especially when you're running at 60 miles

[00:08:59] an hour. You don't have the time to reinvent a wheel. So our focus has been to build embedded

[00:09:06] platforms with all of the neuroscience learnings already embedded. So the amount of training you

[00:09:14] have to do is not really training, but guardrails. If you are in the soup business and you say,

[00:09:20] the kind of things I want to talk about are flavors and ingredients and freshness,

[00:09:25] you give some guardrails and automatically, knowing your audience, the system adjusts and

[00:09:30] automatically prepares for it. Or you may be a cookie maker and you may say, you know, I really

[00:09:36] want to talk about the kind of cookies that we make that do not have, you know, artificial ingredients

[00:09:42] and they're super healthy, they're keto friendly. I want you to limit it to that, but talk about

[00:09:48] deliciousness. So as a small and medium-sized business owner, you should only provide guardrails

[00:09:55] that differentiate and talk about your business and not worry about anything else. No one has the time

[00:10:02] when you're baking cookies to think about what neuroscience principles are. You don't have the

[00:10:06] time and you shouldn't have to because enough scientists and graduate students and market researchers

[00:10:12] have spent decades working through all of that. It is what I call the building blocks of human knowledge

[00:10:20] that should not be jettisoned in the arrogance of an LLM. Just because you have a large language model

[00:10:28] that supposedly has read everything and its brother, you can't glibly say we don't need mathematics or

[00:10:36] science or physics or law or musicology anymore. It would be dumb to say that. And I call it a lot of

[00:10:44] hubris, right? A little bit of humility where we say that the domain knowledge that humans have worked

[00:10:51] so hard to create can be intelligently embedded so that the algorithms serve us and don't walk in

[00:11:01] with the hubris that they replace us. How arrogant, right? Long ago in my career, I used to design

[00:11:08] weapons and I stopped designing them because I said to myself, you know, I was a scientist at GE's R&D a long

[00:11:13] time ago. I said, I will never build a gun to humanity's head and pull the trigger. I don't

[00:11:19] question people who do that and they do that for a living and I appreciate that. But I decided I won't

[00:11:23] do that. So when we built our platform, we decided we are not going to build an AI platform that

[00:11:30] supposedly will replace humanity. It's going back to being a weapons designer that I didn't want to do.

[00:11:36] So we wanted to create a platform that will empower us. So think about it, Dave, right?

[00:11:43] If you have, I always thought I was a musician. My mother tried very hard to make one of me and I think

[00:11:49] she failed miserably, right? I know I have a lot of locked music in my mind. I don't have the talent

[00:11:56] to get it out. Yet through the intelligent use of algorithms, we are able to unlock the musician in each

[00:12:04] one of us, the artist in each one of us, the writer in each one of us. I really view our platforms as

[00:12:12] empowering the latent genius within. And so I don't want the mom and shop owner of a retail business or a

[00:12:23] small and medium-sized business to recreate the lessons of neuroscience. Please do what you're awesome

[00:12:29] and how you want to disrupt and change and transform your little community and the world.

[00:12:35] The lessons of neuroscience that are important to you as you market and message and price and optimize

[00:12:42] your product, which you've never had time to do, now the platform will do for you. And therefore,

[00:12:49] empowering you to be the differentiated business owner that you are.

[00:12:54] Well, you're thinking it's very similar to mine because my experience so far, significantly more

[00:13:00] limited is, but the use cases where I apply particularly generative AI to make something

[00:13:07] better that's human created is where I'm having success and enhancing it and working it as a

[00:13:12] partner where I try and replace something. It's always weaker. Your conversation leads me to two key

[00:13:18] questions that I want to understand. It seems like one of the things you've done really well here,

[00:13:22] and I want to make sure I'm understanding this right, is all of the work of kind of profiling

[00:13:28] customers and creating customer personas and trying to design targets. That's what you're talking

[00:13:33] about neuroscience delivering is, is if I'm trying to think about a typical buyer who is between 18 and 25

[00:13:42] female, I can, I can now work with an AI that is trained to speak that style of, of, of language for

[00:13:49] better. Is that framing sensible? Am I missing detail?

[00:13:54] It's very sensible, but it is one part of the story, right? Consumer understanding,

[00:13:58] deepening the consumer understanding and therefore knowing how to evoke desire, how to price right,

[00:14:06] how to promote right is one part of the story. The second part of the more interesting story

[00:14:11] is creating products for them. So yeah, you know your consumer is 18, is health conscious,

[00:14:17] is running around crazy. So what products do you create for them? Well, you've created something,

[00:14:24] can you extend it? Can you win their hearts and minds with it? What innovations should you create?

[00:14:30] So I have a little over 90 patterns, so I am mostly a tinkerer, like most of us, we are tinkerer,

[00:14:36] so we just mess around with ideas and thoughts. So we took all of our innovative abilities and had our

[00:14:44] algorithms innovate, right? So, and that's where it is fabulous fun working with the algorithms to

[00:14:52] innovate because we do some interesting things. Knowing the consumer and also knowing what their

[00:14:58] non-conscious needs are. Because you see, the big problem in market research, Dave, is this,

[00:15:04] you go to the consumer and you do focus groups and surveys, tell me what you want, tell me what you

[00:15:08] want. Steve Jobs used to say, don't ask your consumer what they want, they don't know what they want.

[00:15:14] However, when you show the consumer a fabulous product, they say, oh my God, thank you for making

[00:15:20] it. That's exactly what I wanted. But how could the consumer, who could not articulate what they

[00:15:25] wanted, somehow recognize what they needed? How funny is that? Well, that's because most of our

[00:15:32] decision making, 95% of it is in the non-conscious, right? We don't have a window into our non-conscious

[00:15:40] data. And we have the world's largest collection of non-conscious data we look through for innovations.

[00:15:48] You may say, that's a very strange word. What is non-conscious data? Well, we'll get to that in a

[00:15:53] second. But the idea is looking through and finding what the unarticulated hidden needs of a consumer

[00:16:02] may be. And then figuring out what product innovations or even product positioning may help

[00:16:10] the consumer appreciate the product the small business made, right? If you're a small and medium

[00:16:16] sized business owner, you don't have hundreds of millions of dollars to go advertise. You just don't.

[00:16:22] On the other hand, if you strategically either extended your product a little bit,

[00:16:29] innovated a little bit, or strategically positioned your product a little differently, boom, everything

[00:16:36] clicks. And the consumer says, that's exactly what I needed. How did you know I needed that,

[00:16:40] right? So what we do, our algorithms innovate. But we do something very interesting when you innovate.

[00:16:47] Right? So for instance, Dave, you know, if you are in a company, and let's say I report to you,

[00:16:52] and you say, Pradeep, I want you to bring your team and innovate today on, I don't know,

[00:16:57] food products, right? And then you say, by the way, I'm going to have a speaker to inspire you all.

[00:17:03] And you bring some speaker, I don't know, from shoe polish or whatever it is, is going to talk about

[00:17:07] what they did. And then you say, okay, boys and girls, be inspired by the speaker and go innovate for me.

[00:17:12] That is to say, we as human beings have extraordinary abilities to be quote unquote,

[00:17:20] inspired by innovations in different domains, and find a way to translate them, and find analogies in

[00:17:28] our own domain. And we have taught the algorithms how to do that. And you would be stunned as to

[00:17:37] what comes out of it. Are you still there, Dave?

[00:17:40] I am. I've been listening. So I'm curious, I would feel like this is an area that could be

[00:17:46] fraught with bias. And this is one of those areas where, you know, and let's think about two men

[00:17:51] talking to one. Yeah, right. So that's at least there's an inherent bias just alone in that.

[00:17:55] I'm curious, wouldn't the how does this not amplify bias in a way by trying to apply the various and how

[00:18:04] are you addressing the potential bias of humans training algorithms to try and think this way?

[00:18:10] How does it address that?

[00:18:12] Yeah. So the interesting thing is this, right? The algorithms inherently

[00:18:21] work through the intelligent use of how we have told them they should think. That is,

[00:18:28] you could either let the algorithms meander through like cows grazing in a field and come up with

[00:18:33] whatever they want to do, in which case you would find that some cows automatically go to one side of

[00:18:40] the field and some others don't. And you say, my gosh, you know, there are biases in here that you have

[00:18:45] introduced into this thing. What we do is this. When you look for innovations, first you take the target

[00:18:52] group that you're interested in. If it is young women who are in Manhattan, who are in the ages of 18

[00:18:59] through 30 in their first job running around, then the algorithms automatically are guided by

[00:19:07] neuroscience underpinnings of what these young women generally, the way their brains are wired and they

[00:19:15] work. But then what we do today, we look at what we call memory structures. What are memory structures?

[00:19:22] It's very simple. There are things like episodic memory, procedural memory, temporal memory. You know,

[00:19:29] I have not been to your house, Dave, but I would guess that you have maybe a bathroom, a kitchen, a living

[00:19:35] room, etc. in your home. I've never been there. That is to say the architecture of memory, while each one of us

[00:19:43] may carry different memories, the architecture of them are similar because we are all human beings. So what we do,

[00:19:51] we look for memory structures in a target audience. And if there is a young woman that typically loves to go to an

[00:19:59] orange theory or a yoga studio or whatever, some of what she desires, wants, needs, seeks, may be governed by those

[00:20:09] places she visits and non-consciously she's thinking about them as she runs through the day. So once you understand the

[00:20:17] her memory structures, they guide the algorithms into starting to think about what may be her needs.

[00:20:25] So the funny thing is, algorithmically innovating, contrary to what we generally think, is more

[00:20:34] unbiased. It is not governed by what Dave or Pradeep think about what this young woman does, needs, and what

[00:20:42] she should want or should need. But rather, it is more unbiased. It is the closest you would ever get to a

[00:20:50] champion of that particular demography. So the funny thing is, if we used AI, right, it actually would be

[00:20:59] the most fabulous representative of the demography you see and can actively walk away from biases that

[00:21:09] cause you to think what they may or may not need, right? Because we all impose our biases of what somebody

[00:21:17] may or may not need on what we innovate for them. And I would honestly say, the trillion dollar shopper

[00:21:23] is a woman, trillion dollar shopper, yet most products for her are made by guys. Messaging for her, written

[00:21:32] by guys. Marketing for her, done by guys. What a crying shame, right? So you find that the use of

[00:21:40] generative AI can actually unbiased a lot of things and make it a little closer to what the demographic

[00:21:48] wants. As an example of the innovation thing that I was talking to you about, the very large company's

[00:21:56] chief product officer was on the phone with me and I was doing a demo of the platform. And he said,

[00:22:02] Pradeep, I'm going to give you a challenge. I want you to innovate in food products for me,

[00:22:08] right? Food products and packaging. And I said, great, I can have the algorithm run through it.

[00:22:13] Tell me what category that you wanted to be inspired by. He said, yeah, I'll give you a challenge.

[00:22:18] I want to be inspired by deodorants. We have to innovate in food products and packaging,

[00:22:27] but we need to be inspired by deodorants. Now, I am a creative guy, but I don't know what these

[00:22:33] algorithms are going to come up with in real time, right? And there was no backing off this demo.

[00:22:39] And he's sitting there and we are hitting the keystroke. The algorithm comes back with

[00:22:46] a food package with an outside scratch and sniff thing, where if you open the flap, it will smell

[00:22:56] like the food when it is fully prepared. And he was stunned because so far in the world of food,

[00:23:04] food packages, prepared food, et cetera, we only showcase delicious food on the cover, right? On

[00:23:11] the package. It's mouthwatering good, we will say. But in the gustatory complex in the brain,

[00:23:18] flavor and fragrance are related. And that is why during COVID, when you couldn't smell the food,

[00:23:24] you didn't taste anything, right? So if you can smell the prepared food on a flap on the outside

[00:23:31] of the food package, your mind already is thinking about, oh my God, how delicious would it be to eat

[00:23:38] it, right? I was stunned as well. And this is where the working with the algorithms is just gorgeous

[00:23:47] because as creative people, I love the way you said that you found greater pleasure in working with it

[00:23:53] not in trying to have it replace you. And I fully concur as a creative sidekick, an aide,

[00:24:01] a guide, an apprentice, or just a co-worker. It's extraordinary what we are able to do and it

[00:24:09] just opens up our mind. And it's just a fabulous thing. And I really believe that large companies,

[00:24:17] they have hundreds of employees sitting around thinking of how they could innovate. Small companies

[00:24:24] don't. Medium-sized businesses don't. And in addition, think about this. Let's say you're a small

[00:24:30] and medium-sized business and you have come up with a lovely innovation. The first thing as a CEO,

[00:24:35] you worry, will somebody copy it? Will a big company run away with it when I put it in the marketplace?

[00:24:41] So we have created something that automatically generates a provisional patent, right? So anytime

[00:24:50] you innovate, it automatically generates a provisional patent along with claims, the whole

[00:24:57] kit and caboodle. So you could take it to a paralegal or somebody who is cost-effective

[00:25:05] and they can dust it up a little bit and get it submitted. For small businesses, you can't afford

[00:25:12] to pay $25,000 every time you submit a patent. On the other hand, if you can get all the way there

[00:25:17] and just get the last bit dusted up and made professional, you get a chance to protect your IP.

[00:25:23] You do it cost-effectively. And instead of just generating content factories, you generate IP

[00:25:31] factories and a small business can suddenly protect itself and its innovations.

[00:25:37] Well, you have given me an incredible use case to end on. So that is exactly where I want to wrap

[00:25:41] it. Dr. A.K. Pradeep is a pioneer in the field of consumer neuroscience and the CEO of sensory.ai,

[00:25:48] where he blends neuroscience with generative AI to help global brands unlock the power

[00:25:53] of the non-conscious mind. With over 90 patents and a track record of transforming how businesses

[00:25:58] connect with consumer, Dr. Pradeep's insights redefine the future of AI-driven marketing and

[00:26:04] innovation. Dr. Pradeep, if people are interested in getting more information, where can they go?

[00:26:09] Two places, pradeep at sensory.ai, P-R-A-D-E-E-P at sensory.ai. Or you could just buy my book,

[00:26:18] Neuro AI on Amazon. There it is.

[00:26:23] The easiest way to do so. Well, thanks so much for joining me today.

[00:26:26] My pleasure, Dave. My pleasure. Thank you very much. Cheers.

[00:26:51] With a no-code control approach, CoreView revolutionizes your Microsoft 365 administration.

[00:26:57] This powerful platform enables automatic reporting and remediation, ensuring optimal performance

[00:27:03] and security. The best part? You achieve this high level of service without the need for a large

[00:27:09] workforce, allowing you to focus on growing your business through efficiency. Want to know more?

[00:27:15] Visit coreview.com slash MSP and find out more.

[00:27:21] The Business of Tech is written and produced by me, Dave Sobel, under ethics guidelines,

[00:27:26] posted at businessof.tech. If you like the content, please make sure to hit that like button

[00:27:32] and follow or subscribe. It's free and easy and the best way to support the show and help us grow.

[00:27:39] You can also check out our Patreon, where you can join the Business of Tech community

[00:27:43] at patreon.com slash MSP radio or buy our Why Do We Care merch at businessof.tech. Finally,

[00:27:53] if you're interested in advertising on this show, visit MSP radio.com slash engage. Once again,

[00:28:00] thanks for listening to me and I will talk to you again on our next episode of the Business of Tech.

[00:28:08] Part of the MSP Radio Network.